运筹与管理2024，Vol.33Issue(12)：17-23,7.DOI:10.12005/orms.2024.0381
考虑质量损失的再制造产品线多目标联赛竞争优化
Multi-objective League Championship Optimization for Remanufactured Product Line Considering Quality Loss
摘要
Abstract
Promoting the efficient recycling of resources and vigorously developing the remanufacturing industry are significant strategies of products made in China 2025.Improving the comprehensive utilization of waste electrical products can accelerate the transformation and upgradation of product lines.In practice,the quality of remanufactured products is often considered lower than that of new products,which has restricted the develop-ment of the remanufacturing industry.Recycling the parts with less quality loss is one of the ways to ensure the quality of remanufactured products,but the recycling cost is high;on the contrary,although recycling the parts with more quality loss can reduce the recycling cost,it increases the remanufacturing cost.The government adopts an appropriate subsidy strategy for remanufacturers according to their resource utilization rate,so as to reduce the production cost of remanufactured products.There are three handling options for recycled parts:direct use with simple treatment,repair,and upgradation.In order to maximize the value of remanufacturing and mini-mize the cost of remanufactured products,it is necessary to develop a decision-making method to determine the best option of recycled parts and optimize the design of the remanufactured product line,so as to meet the hetero-geneous needs of customers in each market segment. Based on different quality characteristics of recycled parts,this study proposes a multi-objective program-ming model of remanufactured product line considering quality loss.Firstly,recycled parts usually have three quality characteristics,including the-larger-the-better characteristic,the-smaller-the better characteristic,and the nominal characteristic.Each quality characteristic has a different degree of impact on remanufactured product quality.After the loss function of each quality characteristic is handled as dimensionless,and the quality loss caused by all quality characteristics is accumulated,this study establishes multivariate quality loss function model.Secondly,analyzing the trend that recycling cost decreases along with an increase in quality loss of recy-cled parts,we define the relationship between costs using a decreasing downward convex function based on the statistical analysis of the recycling valuation platform for used parts.Thirdly,in addition to the recycling cost,the total remanufacturing cost also includes the costs of repair,upgradation,purchase of new parts,and assembly.An optimization model is constructed to minimize the quality loss of recycled parts and minimize the total remanufacturing cost,which determines the product launch,configuration,and handling method of recycled parts. A multi-objective league competition algorithm(MOLCA)is designed to solve this model with conflicting objectives.MOLCA simulates multiple teams competing in a tournament for several seasons(each season contains several rounds).According to the league schedule for each round,teams play in pairs,and winners are determined based on the strength of each team.After each round of the tournament,each team should update its formation and design a new play style based on its own competition experience and the performance of opponent team in the next round.The tournament will last for several seasons,and eventually the best team will be the champion.The team formation is updated in MOLCA by simulating the SWOT analysis process.This updating not only focuses on the formation toward the strong team but also pays attention to one away from the weak team.According to the Pareto dominance relationship,the proposed algorithm judges the winner and loser,and finds the Pareto optimal solutions for solving the multi-objective optimization model in this paper. The extensive experiments are conducted to compare MOLCA with Non-dominated Sorting Genetic Algorithm II(NSGA-II).From the perspectives of convergence and diversity of Pareto optimal solutions,the advantages and disadvantages of the Pareto solution are measured by three metrics:coverage metric,spacing metric,and maximum spread metric.Eight groups of experiments are designed,each of which only changes one key parame-ter to test its impact on the two objective function values.The other parameters are constant or random integers between the minimum and maximum.Ten instances are randomly generated for each set of key parameters.The average values of the three metrics are calculated to test the influence of each key parameter on the performance of MOLCA and NSGA-II.In the end,the numerical experimental results show that the solutions of MOLCA are more dominant,more uniform in distribution and more convergent than those of NSGA-II within the same runtime.Consequently,MOLCA has superior optimization performance and solution quality.关键词
再制造/多目标规划/质量损失/联赛竞争算法Key words
remanufacturing/multi-objective program/quality loss/league competition algorithm分类
管理科学引用本文复制引用
柳春锋,王德丽,王居凤..考虑质量损失的再制造产品线多目标联赛竞争优化[J].运筹与管理,2024,33(12):17-23,7.基金项目
教育部人文社会科学研究规划基金项目(21YJA630065) （21YJA630065）
